[15856] in Kerberos_V5_Development
Re: krb5 libraries and circular dependencies
daemon@ATHENA.MIT.EDU (Sam Hartman)
Fri May 28 15:32:42 2010
From: Sam Hartman <hartmans@mit.edu>
To: Tom Yu <tlyu@mit.edu>
Date: Fri, 28 May 2010 15:32:07 -0400
In-Reply-To: <ldvk4qnompm.fsf@cathode-dark-space.mit.edu> (Tom Yu's message of
"Fri, 28 May 2010 15:22:45 -0400")
Message-ID: <tsld3wfvn48.fsf@mit.edu>
MIME-Version: 1.0
Cc: krbdev@mit.edu, Nicolas Williams <nicolas.williams@oracle.com>
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Errors-To: krbdev-bounces@mit.edu
>>>>> "Tom" == Tom Yu <tlyu@MIT.EDU> writes:
Tom> Nicolas Williams <Nicolas.Williams@oracle.com> writes:
>> I would definitely not be concerned about loading the wrong
>> libkrb5. Distributions can easily ensure that that cannot happen
>> with the bits they ship, and users should know not to mix and
>> match plugins built against one libkrb5 with another -- DLL hell
>> can result if they do.
Tom> I disagree that we should be unconcerned about that. I have
Tom> heard plenty of vitriol from developers and sysadmins who
Tom> experienced trouble from the Unix/Linux equivalent of DLL hell
Tom> when installing software that uses plugins. If the OS vendor
Tom> or software packager never makes any mistakes, and if the
Tom> sysadmin never installs a more recent version of some software
Tom> built from source, then _maybe_ it is possible to ensure that
Tom> loading the wrong library "cannot happen".
If the krb5 build system didn't use rpath, I think it would be fairly
unlikely that a plugin would get a different libkrb5 than the
application. However, since rpath is used, I guess it is reasonably
easy to have this happen.
--Sam
_______________________________________________
krbdev mailing list krbdev@mit.edu
https://mailman.mit.edu/mailman/listinfo/krbdev
